Machine learning device, control system, and machine learning

1. A machine learning device that performs machine learning of optimizing first coefficients of a filter provided in a motor controller that controls a motor and second coefficients of a velocity feedforward unit of a servo control unit provided in the motor controller on the basis of an evaluation function which is a function of measurement information after acceleration and deceleration by an external measuring instrument provided outside the motor controller, a position command input to the motor controller, and a position error which is a difference between the position command value and feedback position detection value from a detector of the servo control unit.

2. The machine learning device according to claim 1 , wherein the measurement information of the external measuring instrument includes at least one of a position, a velocity, and an acceleration.

3. The machine learning device according to claim 1 , wherein the servo control unit includes at least one of a position feedback loop and a velocity feedback loop, and the filter is provided outside the position feedback loop or the velocity feedback loop.

4. The machine learning device according to claim 1 , wherein the motor control unit has a feedback loop and the measurement information of the external measuring instrument is not used for feedback control of the feedback loop.

5. The machine learning device according to claim 1 , wherein the external measuring instrument is detached after adjustment of the filter by machine learning.

6. The machine learning device according to claim 1 , wherein the filter is an IIR filter of the second order or higher.

7. The machine learning device according to claim 1 , further comprising: a state information acquisition unit that acquires state information including the measurement information, the position command, the position error, the first coefficients of the filter, and the second coefficients of the velocity feedforward unit;

an action information output unit that outputs action information including adjustment information of the first and second coefficients included in the state information to the filter and the velocity feedforward unit;

a reward output unit that outputs a reward value of reinforcement learning using a value function which is a function of the measurement information, the position command, and the position error; and

a value function updating unit that updates an action value function on the basis of the reward value output by the reward output unit, the state information, and the action information.

8. The machine learning device according to claim 7 , further comprising: an optimization action information output unit that outputs adjustment information of the first and second coefficients on the basis of the value function updated by the value function updating unit.

9. A control system comprising: a motor controller that includes the machine learning device according to claim 1 , a motor, a servo control unit that includes a velocity feedforward unit, and a filter, the motor controller controlling the motor; and

an external measuring instrument provided outside the motor controller.

10. A machine learning method of a machine learning device, comprising: acquiring first coefficients of a filter provided in a motor controller that controls a motor, second coefficients of a velocity feedforward unit, measurement information after acceleration and deceleration by an external measuring instrument provided outside the motor controller, a position command input to the motor controller, and a position error which is a difference between the position command value and feedback position detection value; and performing machine learning of optimizing the first coefficients and the second coefficients on the basis of an evaluation function which is a function of the measurement information, the position command, and the position error.